Allele age

Estimating allele age based on the allele's frequency is based on the fact that alleles in high frequency are older than alleles in low frequency (assuming the absence of selection). Of course, many alleles of interest are under some type of selection. Because alleles that are under positive selection can rise to high frequency very quickly, it is important to understand the mechanisms that underlie allele frequency change, such as natural selection, gene flow, genetic drift, and mutation.

Estimating allele age based on intra-allelic variation is based on the fact that with every generation, linkage with other alleles (linkage disequilibrium) is disrupted by recombination and new variation in linkage is created via new mutations. The analysis of intra-allelic variation to assess allele age depends on coalescent theory. There are two different approaches that can be used to analyze allele age based on intra-allelic variation. First, a phylogenetics approach extrapolates an allele's age by reconstructing a gene tree and dating the root of the tree. This approach is best when analyzing ancient, as opposed to recent, mutations. Second, a population genetics approach estimates allele age by using mutation, recombination, and demography models instead of a gene tree. This type of approach is best for analyzing recent mutations.

Recently, Albers and McVean (2018) proposed a non-parametric method to estimate the age of an allele, using probabilistic, coalescent-based models of mutation and recombination. Specifically, their method infers the time to the most recent common ancestor (TMRCA) between hundreds or thousands of chromosomal sequence (haplotype) pairs. This information is then combined using a composite likelihood approach to obtain an estimate of the time of mutation at a single locus. This methodology was applied to more than 16 million variants in the human genome, using data from the 1000 Genomes Project and the Simons Genome Diversity Project, to generate the atlas of variant age.

History

Population geneticists, Motoo Kimura and Tomoko Ohta, were the first to analyze the association between an allele's frequency and its age in the 1970s. They showed that the age of a neutral allele can be estimated (assuming a large, randomly mating population) by

${\displaystyle E(t_1)=(-2p)/(1-p)\ln (p)}$

Where ${\displaystyle p}$ represents the allele frequency and ${\displaystyle t_1}$ is the expected age, measured in units of 2N generations.

More recent studies, however, have focused on the analysis of intra-allelic variation. In 1990, Jean-Louis Serre and his team were the first to assess allele age by analyzing intra-allelic variation. Using a sample of 240 French families, they surveyed two restriction fragment length polymorphisms (RFLP) sites (E1 and E2) that are closely linked to an allele (ΔF508) at the cystic fibrosis locus (CFTR). Recombination theory allows for the calculation of x(t), the expected frequency of E2 in association with the allele ΔF508 in generation t, and y, the frequency of E2 on chromosomes without the ΔF508 allele. The recombination rate, c, is assumed to be known, and so the allele age can be calculated as an estimate of t.

${\displaystyle t=\frac{1}{\ln (1-c)}\ln \frac{x(t)-y}{1-y}}$

Although Serre et al. (1990) were the first to employ this method, it became increasingly popular after the Risch et al. study in 1995, which analyzed alleles in an Ashkenazi Jewish population.